Development, Feasibility, Acceptability, and Impact of a Positive Psychology Journaling Intervention to Support Addiction Recovery

Abstract

Recovery from substance use disorders is an arduous, lengthy process, yet there exist few interventions specifically designed to make recovery more reinforcing. This single-group mixed-methods study describes the development, feasibility, acceptability, and impact of a novel journaling intervention that combines aspects of positive psychology and behavioral activation to make recovery more reinforcing. We taught the journaling to women in residential treatment for substance use disorders during eight group sessions, after which participants journaled independently for 2 weeks. The journaling practice was found to be feasible and acceptable. Quantitative data showed improvement in a range of well-being, recovery, and mental health–related factors. Qualitative data showed that the intervention helped participants to recognize what was positive about recovery, to achieve meaningful short-term goals, and to experience a sense of optimism and pride in their accomplishments.

Researchers have argued that addiction treatment should do more than reduce or eliminate substance use. Treatment should also help clients build a pleasurable, satisfying life in recovery (DiClemente, 2003; Marlatt & Gordon, 1980; McKay, 2017; Witkiewitz et al., 2020). To address this need, many researchers are leveraging the principles of positive reinforcement (Higgins et al., 2004). Positive reinforcement, and its attending pleasurable rewards, provide a useful framework for understanding addictive behavior. Reinforcement explains why some people repeat the use of substances after first initiation and why others continue the use of substances despite negative consequences. The principle of reinforcement can also be applied to recovery; if recovery becomes more reinforcing than active addiction, people will do the hard work needed to maintain it (Laudet, 2011; McKay, 2017; Rothman, 2000; Rothman et al., 2016). And yet, this task is formidable. For one thing, individuals in early recovery confront the negative social and psychological consequences of their addictive behavior (Forcehimes et al., 2007; Klingemann & Gmel, 2001; Miller et al., 1995); for another, physiological changes in the brain suppress positive affect while increasing stress reactivity (Ahmed et al., 2002; Koob, 2008). These conditions persist during early recovery, making it more challenging to access positive experiences.

While some existing treatments have components that are designed to make recovery more reinforcing, the independent effect of such reinforcement components is unknown. In addition, interventions explicitly designed to make recovery more reinforcing have shown promise but have not been combined to enhance impact. The current study fills this gap by reporting preliminary results on the development, feasibility, acceptability, and impact of a novel journaling intervention, “Positive Peer Journaling” (PPJ). PPJ combines elements from positive psychology and behavioral activation to make life in recovery more reinforcing. Our review of the relevant research focuses on existing interventions designed to make life in recovery more reinforcing as well as extant research on journaling in general and in the addictions field.

Interventions Designed to Make Life in Recovery More Reinforcing

Components in well-regarded, evidence-based treatments are designed to make recovery more reinforcing, although scant research has evaluated the independent effect of these components. The community reinforcement approach (CRA) leverages positive reinforcement as its primary mechanism of change (Meyers et al., 2011), targeting life areas where clients report being unhappy and encouraging engagement in enjoyable substance-free activities (Meyers et al., 2011). Some aspects of the CRA have been studied for their independent effects, such as behavioral couples therapy, mutual-aid participation, and employment programs; but research specifically focused on CRA’s aspects addressing happiness, enjoyment, hobbies, and recreation have not been evaluated independently (McKay, 2017).

Project MATCH (Project MATCH Research Group, 1993), the largest alcohol treatment trial of its time (Longabaugh & Wirtz, 2001), provides another example of interventions that touch on, but do not explicitly study, happiness in recovery. Project MATCH’s cognitive-behavioral treatment manual featured a session on increasing pleasant activities in recovery (Kadden et al., 2003), and the 12-step facilitation manual featured a session on hobbies and sober living (Nowinski et al., 1992). However, these modules were elective, to be used at the counselor’s discretion, and were not studied for their independent effects, especially considering that one researcher close to the project recalled that they were rarely used (K. Carroll, personal communication, December 23, 2011).

Two sets of studies designed to make addiction recovery more reinforcing, behavioral activation and positive psychological interventions, might be particularly effective if combined. Behavioral activation for addiction is present in two interventions: the Life Enhancement Treatment for Substance Use (LETS ACT, Daughters et al., 2008, 2017; Magidson et al., 2011) and the Substance Free Activity Session (SFAS, Meshesha et al., 2020). Both interventions involve working with individuals with substance use disorders to plan rewarding, sober activities that flow from clients’ values. These studies have shown a range of benefits including greater enjoyment of activities and lower levels of depression and anxiety (Daughters et al., 2008), better treatment retention and higher activity levels (Magidson et al., 2011), better abstinence rates at 3, 6, and 12 months after treatment (Daughters et al., 2017), engagement in lower proportions of substance-related activities, and lower levels of alcohol demand (Meshesha et al., 2020).

The second promising line of research that focuses on making recovery more reinforcing is positive psychology. Positive psychology is the scientific study of subjective individual experiences, personal traits, and institutions associated with a higher quality of life and a life well lived (Seligman & Csikszentmihalyi, 2000). From its inception, positive psychology has tested interventions to improve mood and well-being; meta-analyses within the field have reported small (Bolier et al., 2013) and medium effect (Sin & Lyubomirsky, 2009) sizes for favorable impact on well-being and depression. Scholars have begun applying positive psychological interventions to addiction treatment and recovery (Krentzman, 2013). These interventions have involved gratitude, identification of personal strengths, acts of kindness, and savoring positive events (Akhtar & Boniwell, 2010; Carrico et al., 2019; Hoeppner, Schick, et al., 2019; Kahler et al., 2015; Khazaei et al., 2017; Krentzman et al., 2015; Parker et al., 2020). This research has shown increases in positive affect (Akhtar & Boniwell, 2010; Carrico, Gómez, et al., 2015; Carrico et al., 2019; Hoeppner, Schick, et al., 2019; Krentzman et al., 2015), increases in flourishing (Parker et al., 2020), and decreases in addictive behavior (Akhtar & Boniwell, 2010; Carrico et al., 2019; Kahler et al., 2015; Khazaei et al., 2017).

While null findings have also been reported, positive psychology for addiction treatment shows promise. One robust finding in the literature is that individuals with addictions have enjoyed such interventions, with participants describing them as enjoyable (Kahler et al., 2013), easy, satisfying, or pleasant (Krentzman et al., 2015), beneficial (Carrico, Gómez, et al., 2015; Krentzman et al., 2015), and worth recommending to others (Carrico, Gómez, et al., 2015; Hoeppner, Hoeppner, et al., 2019). Addiction treatment should be enjoyable; clients will repeat and maintain interventions that are effective and reinforcing. Even if an individual intervention does not have a direct effect on outcomes, but is still enjoyable, it can improve client engagement in treatment overall which will increase exposure to other effective components of treatment (McKay, 2017). In other words, if clients find treatment rewarding, they will be more inclined to enroll in it, remain engaged, and return to it later if needed. Therefore, positive psychological interventions can be useful in and of themselves and in the service of improving treatment retention and treatment re-engagement after dropout.

Combining aspects of behavioral activation with positive psychological interventions could enhance the impact of both approaches on important outcomes such as satisfaction with life in recovery for several reasons. Behavioral activation encourages action that aligns with values, such as getting a house chore done, filling out job applications, or taking medications on time (see Figure 2’s Home/Housing, Work/Education, and Health headings). Behavioral activation also encourages positive, pleasurable, and rewarding activities, such as arts and crafts and socializing with peers (see Figure 2’s Joy and Social headings). These accomplishments and pleasures could well populate the next day’s Good things that happened and Things that I am grateful for lists. These gratitude lists bring attention to what has gone well, which might encourage clients to repeat the activity in the coming day. In general, positive correlations between gratitude and behavioral activity have been reported (Wood et al., 2008, 2009). This association may be causal given the following pathway: Gratitude improves affect (Dickens, 2017), and positive affect encourages more frequent and more diverse types of activity (Fredrickson, 2004b). Therefore, behavioral activation fosters gratitude, which in turn encourages positive behavior.

Figure 2.

Positive Peer Journaling Example

Note. This is a de-identified composite of participant responses. Grand mean of the number of bullet points for each heading across the sample was 6.4 for Good things that happened; 2.9 for Bad things that happened; 9.6 for Things that I am grateful for; 7.0 for Wishes for others; 1.2 for Work/Education; 1.7 for Home/Housing; 2.3 for Joy; 1.9 for Health; 1.8 for Recovery; 1.4 for Spirituality; 1.2 for Community; 1.7 for Social; 1.2 for Financial; and 0.7 for Amends/Repair. The number of bullet points depicted in this figure reflects these grand means rounded to the nearest integer.

Research on Journaling

Journaling is written self-expression taking numerous forms (Boud, 2001; Hayman et al., 2012) and serving various purposes, including fostering biopsychosocial wellness. Literature on therapeutic benefits of journaling has focused on expressive writing and interactive journaling, with a subset measuring the effects of such journaling on addictive behavior. The earliest research on expressive writing (Pennebaker & Beall, 1986), and much subsequent research (e.g., Meshberg-Cohen, 2009), has focused on writing about the personal experiences of traumatic events. A set of typical instructions invites a participant to “write about your deepest emotions and thoughts about the most traumatic experience in your life” (Meshberg-Cohen, 2009, p. 52) for 20 minutes on 4 consecutive days because releasing negative pent-up emotions is believed to improve health (Pennebaker & Beall, 1986). Several meta-analyses of expressive writing have been conducted. Some have shown that expressive writing improves both physical and psychological health (Frattaroli, 2006; Smyth, 1998). Others have shown less convincing evidence for emotional health, with one study reporting a positive effect on physical health, but not psychological health (Frisina et al., 2004), and another showing no effect of expressive writing on depression (Reinhold et al., 2018). The weakest findings come from a meta-analysis that found no effect of expressive writing on either the physical or psychological health of individuals with cancer diagnoses (Zachariae & O’Toole, 2015).

Some researchers have resisted the notion that, to be beneficial, expressive writing must focus on negative experiences. However, among college students, studies considering the effect of expressive writing on positive topics have shown it to be associated with fewer visits to the student health center, greater subjective well-being, and increased positive affect (Burton & King, 2004; King, 2001; King & Miner, 2000).

Expressive writing has been tested in only three studies of individuals with substance use disorders, the most robust being a randomized controlled trial of 149 women in residential treatment (Meshberg-Cohen, 2009). Mental health symptoms for the treatment group in this study were lower at the 2-week follow up, although at the 4-week follow up there were no differences between treatment and control groups. Results were likewise mixed in a study with expressive writing as one component of an intervention designed to support emotional processing among 23 men who had had sex with men, were HIV-positive, and had used methamphetamines (Carrico, Nation, et al., 2015). Those who got the intervention used drugs less at the 1-month follow up, yet the control group reported less risky sexual behavior and lower rates of HIV-related stress at the 3-month follow up. These mixed effects could have been due to expressive writing or to other intervention components. An uncontrolled study of 48 individuals with severe substance use disorders examined the effect of expressive writing on psychological and physical health (Baikie et al., 2006), with results showing no improvement at a 2-week follow-up, although these findings were based on the 14 people who had been retained. Of relevance to the current study, the participants in these investigations were enthusiastic about the writing exercises (Baikie et al., 2006), found the process helpful (Baikie et al., 2006; Meshberg-Cohen et al., 2014), and would recommend expressive writing to others (Carrico et al., 2015). Similar to positive psychological interventions, clients’ enjoyment of expressive writing suggests that (a) if it works for them, they will do it and (b) expressive writing is another intervention that could make treatment overall more enjoyable and thereby drive up retention, engagement, and effectiveness (McKay, 2017).

Interactive journaling is a type of therapeutic journaling that has been found to support efforts to initiate behavior change. Interactive journaling entails using a booklet resembling a colorful magazine that includes didactic health-related information followed by prompts inviting a client’s written response within the booklet (Miller, 2014; Proctor et al., 2012; The Change Companies, 2019). Interactive journaling has been shown to improve knowledge about alcohol and to bolster willingness to change among individuals convicted of driving under the influence (Scheck et al., 2013). Interactive journaling also has been shown to reduce rates of criminal recidivism among individuals who have substance use disorders (Proctor et al., 2012). Of note, research on interactive journaling has focused on the initiation of behavior change rather than its maintenance; it is seen as a “pretreatment tool” increasing receptivity to treatment and helping individuals “as they make the transition from the precontemplation to the contemplation, preparation, or action stage of change” (Proctor, 2012, p. 324). While one interactive journaling booklet, Wellness & Recovery (The Change Companies, 2018), addresses gratitude, creativity, and spiritualty, scholars have yet to assess the impact of wellness-oriented interactive journals on making recovery more reinforcing.

In sum, while the addiction field has a history of applying reinforcement theory to the maintenance of addiction recovery through evidence-based practices, the specific components designed to make recovery positively reinforcing have not been examined independently. Two approaches to making recovery more reinforcing, behavioral activation and positive psychology, have shown promise and two forms of journaling have the potential to improve health and initiate change. We posit that PPJ, a journaling practice where participants complete both positive psychology and behavioral activation exercises, is a feasible and acceptable means of making recovery more reinforcing and thereby reducing recurrence of substance use.

Objectives of the Present Study

The present study is developmental and descriptive, conducted to document our first work administering PPJ to individuals with substance use disorders. Our aims, then, are to (a) optimize the delivery of PPJ as an adjunctive treatment for substance use disorders, (b) determine the feasibility and acceptability of PPJ treatment and study procedures, (c) obtain preliminary data on the association between PPJ and treatment outcomes, (d) obtain preliminary data on the association between PPJ and the mediators through which PPJ might impact treatment outcomes, and (e) understand in participants’ own words the ways PPJ might work to support their recovery. We are interested in the effects of PPJ on treatment outcomes such as (a) discharge from the host setting on good terms versus leaving at staff request or against staff advice, and (b) recurrence of substance use while in treatment at the host setting. We hypothesize that the main effects of PPJ on these treatment outcomes will be mediated by indicators that life in recovery is becoming more positively reinforcing (as indicated by, for example, increasing satisfaction with life and happiness with recovery) although we do not formally test for mediation in this descriptive study.

Method

Trial Design

We used a sequential, mixed-methods single-group design (Greene et al., 1989) comprised of three phases: (a) a group phase where PPJ was taught over eight 1-hour sessions and where group members journaled together, (b) an independent phase, where participants journaled on their own for 2 weeks, and (c) a semi-structured exit interview, where participants described their experiences being in the study. Quantitative data assessed feasibility, acceptability, change in hypothesized mediators, effect sizes, and treatment outcomes, whereas qualitative data provided participants’ impressions of PPJ.

Setting

The study took place within a private, non-profit residential substance-use-disorder treatment program for women in the Midwestern United States. Typical treatment at the host setting included educational lectures, process group therapy, individual therapy, relational-cultural therapy, trauma-informed care, cognitive behavioral therapy, and motivational interviewing. While the setting used an abstinence-based model of SUD recovery, a single recurrence of substance use was not grounds for dismissal from the program. Women typically resided in the setting for 90 days.

Recruitment

Staff at the host setting provided flyers to prospective participants and offered them the opportunity to sign up for appointments with research staff. At these appointments, research staff screened participants, obtained written informed consent, and administered the baseline questionnaire. See Figure 1 for the CONSORT flowchart of participants.

Figure 1.

CONSORT Flowchart of Participants

Participants

To be included in the study, participants had to be at least 18 years old, meet DSM-V criteria for a past-year substance use disorder, have English literacy sufficient to make short written lists and understand survey questions, have a minimum of 2 weeks of abstinence, complete approximately 2 weeks of treatment at the host setting, accept being audio-recorded, and have a planned date of discharge beyond the date range for the study.

Participants were excluded from this study if they reported significant symptoms related to a co-occurring psychiatric condition in the week prior to the baseline interview. To determine this, we screened for symptoms of internalizing disorders (e.g., depression), externalizing disorders (e.g., impulsivity), and past criminal and violent behaviors (e.g., interpersonal violence) using the Global Appraisal of Individual Needs--Short Screener (GAIN-SS; Dennis et al., 2013). Following the GAIN-SS guidelines, we explained to participants what we meant by “significant” as follows: “problems are considered significant when clients have them for two or more weeks, when the problems keep coming back, when they keep clients from meeting their responsibilities, or when they make clients feel like they can’t go on” (Dennis et al., 2013, p. 12). Four individuals (13% of those screened) were excluded from the study for symptoms identified during this procedure; each stated that they had heard voices in the week before the screening interview and that this problem was significant for them. Participants were also excluded from participation if they were unable to give informed, voluntary consent to participate (n = 1, 3%), or if they were discharged from the host setting after screening but before the start of research study groups (n = 1, 3%).

In addition to the primary aims of the current study, we also wished to collect preliminary information about whether PPJ might have special utility for individuals maintaining recovery in small towns and rural communities. Recovery in rural communities warrants additional attention from scholars given certain distinctive challenges associated with getting and staying sober in small towns. For example, it can be difficult to make new friends who do not use drugs, it can be hard to find reinforcing sober activities, and it can be challenging to distance oneself from one’s former reputation, including any associated stigma and attendant gossip (Krentzman & Glass, 2021). PPJ might help encourage positive recovery behaviors, underscore what is going well with recovery, and optimize reinforcement for recovery in such settings. For this reason, we gave preference for admission to the study to individuals with previous experience living in towns with populations under 10,500, according to the US Census. This allowed us to begin the work of gathering impressions of PPJ for individuals in rural communities; this is work that we plan to continue but is beyond the scope of the current study (n = 9 excluded for this reason, 30%).

Sample Size

We enrolled three cohorts of five women (N = 15) in February, April, and July 2019.

Baseline Data

See Table 1 for baseline characteristics of the sample.

Table 1.

Baseline Demographic and Clinical Characteristics of the Sample

Baseline characteristic	Sample range	Mean (SD) or % yes (n)
Age	24–59	36.8 (9.7)
Race
Black or African American		6.7% (1)
White (not Latina)		66.7% (10)
Native American or Alaskan Native		6.7% (1)
Asian or Pacific Islander		6.7% (1)
Multiracial		13.3% (2)
Years of education	10–17	13.5 (1.8)
Past year household income
Under $15,000		73.3% (11)
Number of children	0–7	2.2 (2.0)
Substance most addicted to:
Alcohol		26.7% (4)
Opiates		20.0% (3)
Cannabis		6.7% (1)
Amphetamines		46.7% (7)
Length of sobriety in days	18–121	47.8 (31.6)
Number of previous treatment episodesa	1–10	5.6 (3.0)
Negative consequences of substance useb	10–30	21.0 (6.7)
Alcoholics Anonymous Affiliation Scalec	2.5–7.25	5.3 (1.4)

Note. N = 15.

^an = 14.

^bScale range: 0–30.

^cScale range: 0–9. We used the scale to assess the depth of affiliation to participant’s self-identified primary mutual-aid group: Alcoholics Anonymous (n = 8), Narcotics Anonymous (n = 5), 12-Step meetings (n = 1), and Wellbriety (n = 1).

Description of Positive Peer Journaling Intervention and Rationale for Components

PPJ was designed and developed by the first author and is a brief, daily, written practice meant to be adjunctive to standard addiction treatment. PPJ serves as a review of the previous day, using positive psychology exercises such as gratitude and acts of kindness, as well as a planner for the coming day, using behavioral activation exercises. PPJ is administered by way of a notebook printed at the university copy center with category headings under which participants make bullet-pointed lists. A single PPJ entry is contained within a double-page spread, where participants review the past day on the left-hand side and plan the coming day on the right-hand side. Left-hand page prompts include Good things that happened, Bad things that happened, Things that I am grateful for, and Wishes for others. Right-hand prompts include Work/Education, Home/Housing, Joy, Health, Recovery, Spirituality, Community, Social, Financial, and Amends/Repair, plus space in the lower right corner for participants to write their own category headings. The PPJ journal contains 60 PPJ entry pages, two bookmarks attached to the journal and printed with suggestions for each heading, 12 Important Life Area worksheets, the Each One, Teach One exercise script, and 12 blank pages for notes. See Figure 2 for an example of a completed PPJ entry. Elements of PPJ, its rationale, how it is taught in group, and ancillary group activities are described below.

Gratitude.

There are two gratitude exercises in PPJ. The Good things that happened prompt was inspired by the Three Good Things exercise from positive psychology (Seligman et al., 2005), although PPJ allows more than three items. We taught this activity first and advised participants to begin a PPJ entry with this exercise. We provided the instruction to make a bullet-pointed list of all the good things that happened in the past 24 hours. We provided the hint that “small things count” and written examples of what constituted a “good thing,” such as the weather, a moment with nature, and spending time with loved ones. The Things that I am grateful for prompt invited participants to express appreciation (Emmons & McCullough, 2003; Emmons & Stern, 2013). We instructed participants to make a bullet-pointed list of all the things they were grateful or thankful for today. The hint for this prompt was “things we usually take for granted” and written examples included recovery, my counselor, and good food. We encouraged participants to include as many items as possible in these two gratitude lists. We discussed how these two lists were similar and different: similar in that they both accessed good things and different in that one emphasized more discrete things that happened recently and the other emphasized longer-lasting aspects of life in general. Such gratitude interventions have been shown to improve mood (Dickens, 2017) and psychological well-being (Davis et al., 2016), and have demonstrated benefit for individuals with alcohol use disorders (Krentzman et al., 2015).

Expressive Writing.

The Bad things that happened prompt asked participants to list negative events that took place over the previous day, drawing on this recounting of negative affect to leverage the benefits of expressive writing. Early work on PPJ involved feedback from stakeholders who believed that a Bad things that happened column would aid individuals in recovery (Krentzman et al., 2018). Such a prompt allows PPJ to represent more authentically the nature of life in general, while also indicating when help might be needed, especially in cases where the same “bad” item recurs. We found in our previous study (Krentzman et al., 2015) that when participants reviewed the day to list good things that happened, they also thought of bad things that happened. We encouraged participants to add to this Bad things that happened list only as potential items occurred to them while filling out other aspects of PPJ. Content in the group session included descriptions of the negativity bias and changes in the brain during addiction that make it hard to feel positive emotion and easier to feel stress. Therefore, we described that there was no need to make the Bad things that happened list include as many items as possible, because we tend to place excessive emphasis on bad things naturally and we are using PPJ to counter this tendency.

Acts of Kindness.

The Wishes for others prompt invites participants to make a list of people having a hard time (e.g., individuals who are ill, vulnerable, suffering, or recovering from surgery) as a way of cultivating compassion and reducing self-centeredness. After making this list, participants choose someone from the list and write their name under the Social prompt, indicating an intention to reach out to that person over the coming day. We encouraged direct expressions of caring with the intention that this might foster positive interpersonal relationships. We encouraged participants to make the Wishes for others list as long as possible in part by searching their written Good things that happened, Bad things that happened, and Things that I am grateful for lists to find people to add to the Wishes for others list. Acts of kindness to others benefit the actor (Curry et al., 2018), and loving-kindness meditations, where individuals extend good wishes to others, have been shown to offer a range of psychosocial benefits (Fredrickson et al., 2008).

Setting Achievable Goals that Align with Values.

This values-clarification exercise (Kirschenbaum, 2013) is influenced by behavioral activation (Lejuez et al., 2011) and behavioral activation for addiction recovery (Daughters et al., 2016). Participants choose a heading on PPJ’s right-hand page and answer a set of questions about what they value about that heading on the Important Life Areas worksheet, where participants note short-term goals to help them realize their vision for that life area. Leveraging principles from SMART goal planning (Doran, 1981) and behavioral activation (Daughters et al., 2016; Lejuez et al., 2011), we taught participants to draft action items on the Important Life Areas worksheet to optimize the potential for success. Then, having learned to write achievable tasks aligning with their values, participants applied this skill when planning their upcoming day. A tip for doing so from previous research on behavioral activation (Daughters et al., 2016; Lejuez et al., 2011) is printed on the bottom of each right-hand page (see Figure 2).

Planning pleasant events and activities.

In the Joy section, participants plan something recreational, enjoyable, and/or pleasant for the upcoming day, offering a sense of balance, prioritizing joy in recovery, and countering the notion that this page is simply a “to-do” list. We strongly encouraged participants to plan at least one pleasurable or pleasant activity each day. Research has shown an inverse relationship between engagement in substance-free reinforcing activity and substance use (Acuff et al., 2019).

The PPJ Group

The purpose of the eight 1-hour group sessions was to teach the components of PPJ and to practice them together. The first author designed the group content, selected the methods of instruction, and led the group sessions. See Table 2 for the content of each group session. A treatment manual for PPJ 2021 and sample PPJ pages are available from the first author.

Table 2.

Therapeutic Content of Positive Peer Journaling (PPJ) Group Sessions

Session	Content
1	Introduction to the group and overview of the journal. Learn, journal, and share Good things that happened entries, with attention to small things. Learn and journal Bad things that happened entries.
2	Learn, journal, and share: Things that I am grateful for, with attention to things usually taken for granted. Discussion: How are Good things that happened and Things that I am grateful for similar/different? Learn, journal, and share: Wishes for Others. Power moves: 1) Look through the Good things that happened, Things that I am grateful for, and Bad things that happened lists to find people to add to the Wishes for others list. 2) Choose a person from Wishes for others and enter their name under Social with the plan to express good wishes to them directly in the next 24 hours. 3) Does anything under Bad things that happened require repair or amends? Enter it under Amends/repair.
3	Bad things that happened: how and why we treat this list differently. Negativity bias and changes to the brain during addiction make it harder to feel positive emotion and easier to feel stress. We want to grow the Good things that happened, Things that I am grateful for, and Wishes for others lists but we don’t need to grow the Bad things that happened list, we naturally do this on our own! Discussion: good reasons for having a Bad things that happened list. Journal and share.
4	Introduction to right hand page: plan the upcoming 24 hours. Discuss each column heading and brainstorm entries for each one. Journal and share.
5	Learn and share: important life areas values exercise. Choose a life area on the right hand page. Complete values exercise for chosen life area by answering questions such as, what makes this life area mean a lot to you? Write short-term goals for this life area so they are observable, measurable, and represent the next small step. Troubleshoot difficulties with journaling thus far.
6	Learn and share: second chosen life area. Review writing short-term goals so they are observable, measurable, and the next small step. Journal and share. Challenge: plan something for the Joy list every day.
7	Model and practice the teach-back exercise, each one teach one. Journal and share. Homework: teach PPJ to someone else using the each one teach one framework.
8	Review each one teach one homework. Journal and share. Celebratory ending ritual allowing each participant to express what they got out of PPJ.

Note. “Journal and share” indicates journaling the components of PPJ learned to that point (as in Figure 1) and sharing aloud only what each person feels comfortable sharing. Homework for Sessions 2–6: If group does not meet the next day, journal independently. Sessions 2–8: Groups begin with a discussion of how PPJ is going so far and a “teach back” review. Sessions 1–7: Groups end with the question, “What is one thing you got out of group today?”

We taught PPJ using a Rogerian, person-centered approach that emphasized unconditional positive regard and empathy. We employed a number of techniques consistent with Motivational Interviewing (Miller & Rollnick, 2012) such as open-ended questions to foster discussion, elicit-provide-elicit when describing new concepts, and an overall emphasis on participant autonomy.

In our previous research using the Three Good Things exercise with individuals in outpatient treatment for alcohol use disorders (Krentzman et al., 2015), we learned that some participants had trouble thinking of three good things per day. Therefore, we applied suggestions from Wong’s (2017) work on the benefits of learning gratitude in groups by having all group members, including the group leader, a staff member from the host setting, and a research assistant, journal together and share aloud what we had written in order to provide ample suggestions for the kinds of items to list. We encouraged participants to add to their lists based on what they heard others share. Leveraging the idea that a participant will learn a new skill best by teaching it, we employed “teach backs” where participants volunteered to review content from the previous session. Along these lines, we provided a final homework assignment, the Each One, Teach One exercise, where participants were asked to reinforce what they had learned by finding someone outside of the group and teaching them how to complete a PPJ entry.

When leading the groups, we kept the sessions focused on learning and practicing PPJ. Group members were encouraged to access their other sources of support (therapists, therapy groups, etc.), rather than the PPJ group, for assistance with challenges and problems unrelated to learning PPJ. Therefore, in PPJ groups, members did not identify themselves by drug of choice, length of sobriety, or recurrence to substance use, and nor did they read aloud their Bad things that happened lists. This provided the added benefit of an egalitarian spirit among all group participants and kept the focus on well-being concepts and learning PPJ.

Approval and Consent

The University of Minnesota’s Institutional Review Board approved the study, #00004619, and written informed consent was obtained from each participant at the screening interview.

Subject Remuneration

Participants were compensated in gift cards to a general merchandise retail store. Participants earned $15 for the baseline interview, $15 for the exit interview, $5 for completing shorter questionnaires administered several times a week, and a $20 bonus for completing the baseline and exit interviews plus 90% of all other study questionnaires. Their maximum compensation was $155.

Iterative Modifications to PPJ and Study Logistics

Following each cohort, research staff analyzed participant responses and made minor developmental adjustments to group content, group schedule, study logistics, measurement strategy, and journal contents. The sequencing of some group content was modified; for example, the Each One, Teach One exercise was initially covered in Session 6 but was moved to Session 7.

From the beginning, we found that eight 1-hour group sessions were sufficient to teach PPJ; following Cohort 1, and to accelerate treatment delivery, we adjusted the timing of the group sessions from 2 times per week to 3 times per week. After each cohort, we examined the results of quantitative instruments for ceiling or floor effects, degree of theoretical relevance, and completion burden. Some instruments were removed; others were added. Modifications to the printed journal made after Cohort 1 included adding the Important Life Areas worksheets, the Each One, Teach One exercise, and blank pages for notes to the bound journal. See the specific notes of Table 3 for the measurement protocol changes.

Table 3.

Within-Person Change from Baseline to Exit Interview

						Assessment of Change
Construct/scale	Cronbach’s α at baseline	Scale range	Baseline mean (SD)	End-of-group mean (SD)	Exit interview mean (SD)	b [95% CI] for slope of time (in days)a	Estimate of change from baseline to exit interview [95% CI]b	p	Effect sizec
Affect
PANAS negative affecta	.87	10–50	13.7 (4.0)	14.9 (6.6)	13.6 (4.8)	−0.00007 [−0.10, 0.10]d	a	.999	−0.003
PANAS activated positive affecta	.73	10–50	39.3 (4.3)	38.3 (8.2)	37.3 (9.2)	−0.03 [−0.16, 0.10]	a	.600	−0.25
PANAS unactivated positive affect (serenity subscale)a	.86	3–15	11.0 (2.0)	10.7 (2.7)	10.5 (2.3)	−0.001 [−0.04, 0.04]	a	.947	−0.23
Well-being
Optimism, LOT Scaleae	.76	0–24	13.6 (4.6)	11.6 (7.9)	15.6 (5.0)	0.05 [−0.07, 0.17]	a	.338	0.38
Satisfaction with Life Scaleaf	.86	5–35	15.0 (6.0)	17.8 (6.9)	19.9 (5.9)	0.11 [0.01, 0.21]	a	.036	0.77
WHO Well-Being Indexb	.83	0–25	13.7 (4.2)	15.3 (4.5)	16.1 (5.2)	b	2.88 [0.04–5.72]	.048	0.46
Gratitude Questionnaireb	.84	6–42	36.9 (5.2)	34.6 (6.7)	36.8 (4.3)	b	−0.37 [−4.11, 3.38]	.832	−0.02
Reward Probability Indexb	.86	20–80	58.2 (7.2)	58.6 (11.6)	61.8 (8.6)	b	5.15 [1.54, 8.76]	.010	0.42
Addiction related
Cravinga	g	0–10	4.5 (2.5)	4.2 (3.6)	3.6 (3.9)	−0.04 [−0.07, 0.004]	a	.074	−0.25
Recovery related
Happy with recoverya	g	0–10	6.7 (2.3)	8.7 (1.6)	9.1 (1.0)	0.04 [0.01, 0.08]	a	.012	1.20
Commitment to sobrietyb	.88	5–30	28.8 (1.8)	28.6 (1.6)	27.5 (3.6)	b	−0.74 [−2.84, 1.36]	.439	−0.40
Commitment to sobrietya	g	0–10	9.5 (0.9)	9.2 (1.0)	8.9 (1.4)	0.002 [−0.02, 0.02]	a	.817	−0.44
Confidence staying sobera	g	0–10	6.4 (2.3)	8.8 (1.5)	9.2 (1.0)	0.06 [0.02, 0.09]	a	.004	1.46
Brief Assessment of Recovery Capitalb	.76	10–60	49.3 (5.7)	50.7 (8.4)	50.5 (5.2)	b	2.44 [−2.06, 6.94]	.251	0.20
Mental health
HADS depressionb	.63	0–21	4.2 (2.6)	4.0 (3.3)	2.7 (2.6)	b	−2.00 [−3.47, −0.53]	.012	−0.52
HADS anxietyb	.84	0–21	7.3 (3.4)	8.2 (4.6)	7.6 (3.9)	b	−0.50 [−2.05, 1.04]	.486	0.08
Perceived Stress Scaleb	.77	0–16	6.8 (3.1)	5.8 (3.1)	6.0 (3.0)	b	−1.39 [−3.84, 1.05]	.238	−0.24
Demoralization Scalebh	.95	0–96	31.4 (13.1)	28.7 (20.0)	24.0 (12.4)	b	−6.95 [−14.49, 0.60]	.066	−0.53

Note. Some scales were administered only at 3 key time points: baseline, end of group, and exit interview. Other scales were administered at the 3 key time points and at every upload, i.e., up to 22 measurement occasions. Administration schedules for each scale are depicted in the superscript notes that follow.

^aWe assessed this variable at every upload unless otherwise noted; thus, we assessed change using multi-level modelling controlling for any cohort differences. The assessment of change column for these variables represents the change in slope for each day throughout the study averaged across cohorts.

^bWe assessed this variable only at 3 key time points and assess change from baseline to exit interview controlling for any cohort differences. The assessment of change column for these variables represents the difference (exit interview minus baseline) averaged across cohorts adjusting for cohort and cohort * time.

^cThese are effect sizes of the difference between baseline and exit interview, calculated using Hedges g_av (Lakens, 2013).

^dThe model did not converge for unstructured covariance matrix (UN) therefore we fit and report a simpler model using variance components (VC) which did converge.

^eFor Cohort 1, we assessed optimism at baseline only, except in the case of one individual who also completed the optimism scale at the exit interview; for Cohort 2, we evaluated optimism at every upload; and for Cohort 3, at 3 key time points.

^fWe assessed satisfaction with life as follows: Cohort 1 at 3 key time points; Cohorts 2 and 3 at every upload.

^gWe measured this construct with a single item.

^hWe assessed demoralization for Cohorts 2 and 3 only.

Measurement

Data Collection

To determine when participants completed PPJ entries and to administer assessment instruments, we employed a procedure we called the “upload.” The upload had participants accessing a Qualtrics survey on an iPad that we provided, photographing their journal entry, uploading the image, and then completing study questionnaires. We taught participants this procedure during the group phase of the study by inviting them to do the upload immediately after completing a PPJ entry during three or four of the eight group sessions so that we could provide immediate asSsistance if needed. During the group phase, we also invited participants to do the upload three or four times on their own upon journaling independently as homework. During the independent phase, participants completed PPJ entries and did the upload on their own. We provided iPads to participants during group; we put the iPads on reserve for participants in the resident assistants’ office for independent use.

Measurement Administration Schedule

We administered some instruments at baseline only. Some occurred at three key time points—at baseline, on the last day of group, and at the exit interview—and others were done at three key points and also at every “upload,” meaning every occasion when a participant uploaded a snapshot of their journal entry (up to 22 times over the course of the study). We wished to collect as much information as possible at every upload because frequent measurement administration would increase statistical power for multi-level models, which would improve the likelihood of detecting change in this small sample. At the same time, we had to limit what we asked at every upload to reduce administration burden. Therefore, we assessed some constructs of interest at three key time points only. See the specific notes in Table 3 for the assessment schedule.

Baseline Survey

The baseline survey collected information about participants’ demographic and clinical characteristics including negative consequences of substance use, using the Short Inventory of Problems–Alcohol and Drugs scale (10 items: 0 = never, 3 = daily or almost daily; Hagman et al., 2009) and their level of mutual aid affiliation, using the Alcoholics Anonymous Affiliation Scale (9 items with varying response formats including 1 = yes, 0 = no; Humphreys et al., 1998). We allowed participants to answer the questions of this scale with relevance to their primary mutual-aid group if it was not Alcoholics Anonymous. We assessed for mutual aid participation because it describes an important attribute of the sample, given the significant impact of mutual aid on recovery (Kelly et al., 2020; White et al., 2020) and because some activities of mutual aid overlap with PPJ, such as taking a daily inventory (as in Step 10: AA World Services, 1953) and practicing gratitude (Krentzman, 2019). In future research, it will be important to determine the effect of mutual-aid participation on PPJ given these areas of overlap. The baseline survey also assessed for baseline levels of hypothesized mediators, itemized below.

Assessment of Feasibility and Acceptability

We assessed feasibility and acceptability by (a) computing rates of early termination from the study; (b) calculating the percentage of groups attended; (c) calculating the percentage of time participants completed an upload based on the number of times they were instructed to do so; (d) calculating a grand mean of the number of bullet-pointed items participants wrote in each PPJ entry; and (e) assessing the degree to which PPJ was viewed as easy, rewarding, and helpful. This was obtained at every upload by asking participants to respond to the prompt: “I would say that this positive peer journaling entry was …” (0 = not easy at all, 10 = extremely easy; 0 = not difficult at all, 10 = extremely difficult; 0 = not pleasant at all, 10 = extremely pleasant; 0 = not satisfying at all, 10 = extremely satisfying; 0 = not helpful at all, 10 = extremely helpful). Participants’ relative effort was assessed by the prompt: “For this Positive Peer Journaling entry, I would say that I put forth this much effort …” (0 = no effort, 10 = very much effort).

Assessment of Impact

We administered several psychometric scales and single-item instruments to assess hypothesized mechanisms of PPJ’s potential impact on treatment outcomes as follows. Affect was assessed with three subscales of the Positive and Negative Affect Schedule (Watson & Clark, 1994): activated positive affect (feeling attentive, enthusiastic; 10 items); negative affect (feeling nervous, distressed; 10 items); and unactivated positive affect, using the serenity subscale (feeling calm, at ease; 3 items: 1 = very slightly or not at all, 5 = extremely). Optimism was assessed with the revised Life Orientation Test (10 items: 0 = I disagree a lot, 4 = I agree a lot; Scheier et al., 1994; Scheier & Carver, 1985). Satisfaction with life was assessed via the Satisfaction with Life Scale (5 items: 1 = strongly disagree, 7 = strongly agree; Diener et al., 1985). Craving was assessed with the higher score between two items: strongest urge to drink and strongest urge to use drugs (0 = no urge whatsoever, 10 = strongest urge I have ever felt). Happiness with recovery was assessed with a single item: “In general, I am happy with my recovery” (0 = not at all true, 10 = extremely true).

Commitment to Sobriety was measured two different ways. At 3 key time points, commitment to sobriety was assessed with the Commitment to Sobriety Scale (5 items: 1 = strongly disagree, 6 = strongly agree; Kelly & Greene, 2014). At every upload, commitment to sobriety was also assessed with a single item from the scale: “I am totally committed to staying off of alcohol/drugs” (0 = not at all true, 10 = extremely true). Confidence in staying sober was assessed by the single item: “I am confident in my ability to abstain from drugs and alcohol” (0 = not at all true, 10 = extremely true). Depression and anxiety were assessed with two subscales of the Hospital Anxiety and Depression Scale (seven items with varying response formats, such as 0 = definitely, 3 = not at all; Zigmond & Snaith, 1983). Stress was measured by the Perceived Stress Scale (four items: 0 = never, 4 = very often; Cohen et al., 1983). Demoralization was assessed by the Demoralization Scale (24 items: 0 = never, 4 = all the time; Kissane et al., 2004). Well-being was assessed with the WHO Well-Being Index (five items: 0 = at no time, 5 = all the time; World Health Organization, 1998). Trait gratitude was assessed by the Gratitude Questionnaire (six items: 1 = strongly disagree, 7 = strongly agree; McCullough et al., 2002). Resources that can be leveraged to support recovery were assessed by the Brief Assessment of Recovery Capital (10 items: 1 = strongly disagree, 6 = strongly agree; Vilsaint et al., 2017). The potential to obtain positive reinforcement from the environment was assessed by the Reward Probability Index (20 items: 1 = strongly disagree, 4 = strongly agree; Carvalho et al., 2011). See Table 3 for scale ranges and Cronbach’s alphas at baseline.

We obtained information about host setting outcomes from host setting treatment records. We then calculated the percentage of enrolled participants who were discharged from the host setting on good terms (i.e., “with staff approval”), as well as the percentage of enrolled participants who did not experience a recurrence of substance use between the first PPJ group session and discharge from the host facility.

We employed semi-structured individual exit interviews to collect qualitative data on the impact of PPJ. Exit interviews took place after the 2-week independent practice. Interview questions focused on participants’ experiences both in the study and in practicing PPJ. Interviews were audio-recorded and transcribed.

Data Analytic Methods

Quantitative

We calculated means and standard deviations or percentages to determine baseline characteristics, treatment outcomes, the rate of PPJ upload completion, the relative difficulty, pleasantness, and helpfulness of—and effort expended on—PPJ, the count of items in each PPJ upload, and total scale or subscale scores or for each measurement instrument.

We fitted two models to assess change based on the frequency of measurement administration. For each instrument assessed at every upload, we fitted a linear mixed model allowing for random variation of intercepts and slopes with respect to time (in days), where time was modeled as a repeated measure within individuals using an unstructured covariance matrix. For each instrument assessed at three key time points, we fitted a linear mixed model with time as a categorical, fixed predictor, modeling repeated measurements within persons using an unstructured covariance matrix. We used the MIXED command in SPSS 25 and maximum likelihood estimation to fit both models, and in both we controlled for cohort differences by including dummy variables for the cohort and for the interaction between time and cohort as fixed predictors.

Qualitative

The first author inductively identified themes in the interview transcripts using qualitative data–analytic techniques described by Braun and Clark (2006). Themes reported herein pertain to the ways PPJ was described as supportive of recovery.

Results

Outcomes and Estimation

Feasibility and Acceptability

Five individuals were discharged from the host setting before study activities ended, including one who nonetheless completed study activities using her smartphone. Three individuals stopped participating in the study after the first week of the independent phase, including one who temporarily relocated to receive a higher intensity of medical care in the second week of this phase and then returned to the host setting. Participants attended between two and eight group sessions (M = 6.3, SD = 2.1), with an average attendance rate of 78.3% (SD = 26.5%). Three individuals were discharged from the host setting during the group phase. Excluding these three, participants attended 100%, 97%, and 75% of group sessions for Cohorts 1, 2, and 3, respectively. The lower rate for Cohort 3 was associated with a group time change from 10:30 am to 8:30 am. The 15 participants completed a total of 196 uploads during the study. During the group phase, participants were invited to upload three or four times during group sessions and an additional three or four times as homework. Percentage of uploads completed during the group phase (the number completed divided by the number of times invited to complete) was 80.1% (SD = 35.7%, range = 0.0%–128.6%; the upper range is over 100% because three participants uploaded more often than the minimum required). These values include one person discharged from the host setting before completing any uploads. Percentage of uploads completed during the independent phase was 43.8% (SD = 33.5%, range 0.0%–100.0%), when participants were invited to upload a minimum of four times and as many as 14 times over this 2-week period. These values include three people discharged from the host setting before completing any uploads during the independent phase. The mean of bullet-pointed items listed per PPJ entry (as in Figure 2) ranged from nine to 104 (M = 39.2, SD = 18.2). Participants found PPJ to be easy (M = 8.6, SD = 1.3), not difficult (M = 1.6, SD = 2.1), pleasant (M = 7.9, SD = 1.9), and satisfying (M = 8.0, SD = 1.9). Participants also put effort into each PPJ entry (M = 8.7, SD = 1.5).

Hypothesized Mediators

See Table 3 for the depictions of change in hypothesized mediators. We observed increases in satisfaction with life (0.11 points per day, effect size .77), well-being (2.88 points from baseline to exit interview, effect size .46), rewards obtained from the environment (5.15 points from baseline to exit interview, effect size .42), happiness with recovery (0.04 points per day, effect size 1.20), and confidence staying sober (0.06 points per day, effect size 1.46). We observed decreases in depression (−2 points from baseline to exit interview, effect size −.052). Optimism, craving, recovery capital, stress, and demoralization did not meet the threshold of p > .05 but showed at least small effect sizes in the direction of improvement. Negative mood, gratitude, and anxiety showed no change based on p value or effect size; activated and unactivated positive affect and commitment to sobriety showed no change based on p value but showed small effect sizes suggesting decay over time.

Treatment Outcomes

Approximately three quarters of the sample (73.3%, n = 11) were discharged from the host setting on good terms; three quarters (73.3%, n = 11) did not use substances from the period between the first PPJ group through discharge from the host setting. Taken together, two thirds (66.7%, n = 10) did not use substances after beginning PPJ treatment while also leaving the host setting on good terms.

Qualitative Results

We conducted individual exit interviews with 11 participants. We did not invite three participants to the exit interview because they had been discharged from the host setting at staff request or against staff advice; discharge on these terms was articulated in the consent form as disengagement from the research study. One participant who had been discharged on good terms was unresponsive to outreach. Interviews ranged from 18.8 to 71.6 minutes in length (M = 41.0, SD = 15.1).

Analysis of interview transcripts revealed three primary themes that depicted the ways PPJ supported recovery according to participants: PPJ promoted a positive view of life in recovery, PPJ helped participants meet short-term goals, and PPJ created awareness of progress in recovery. (See Table 4 for representative quotations.) In terms of encouraging a positive view, PPJ showed participants that, contrary to their expectations, good things outnumbered bad things, there was more to be grateful for than they had realized, and bad things appeared less negative. With respect to meeting short-term goals, PPJ helped participants get things done by offering them a vehicle with which to provide themselves written reminders while also helping them organize larger goals into smaller ones. Finally, both the left and right pages of PPJ helped participants realize their progress in recovery, bolstering a sense of optimism and pride in accomplishment.

Table 4.

How Positive Peer Journaling (PPJ) Supported Recovery

Themes	Example quote(s)
PPJ promoted a positive view of life in recovery
Good things outnumbered bad things	“I just found it interesting that even on days that I thought I was having a bad day, my good things that happened to me always have outweighed my bad side. I don’t think there was one day where my bad side outweighed the good side. So, it really has been interesting for me to see that a lot of my time my life is really good when I think that it’s not so great.”
There was more to be grateful for than realized	“[PPJ] makes you slow down and think, ‘Okay, well, what am I grateful for?’ And it’s like, ‘Oh, well, I’m grateful for all of this,’ and actually I could keep going but there’s not really enough space.”
Bad things appeared less negative	“[PPJ] reduced my bad things, because things weren’t really that bad; it was just something that happened that I was maybe able to turn into a good thing, so they aren’t even a bad thing. … The bad things that happened, that column was more of a thinking moment: I definitely had to think more, ‘Was it really bad or was it a good learning opportunity?’ I started as we progressed through the journaling—I noticed without doing it on purpose—that I started to twist [those] bad things into … a thinking opportunity.”
PPJ helped participants meet short-term goals
I would usually forget, but PPJ helped me to remember	“My first example would be like my doctor’s appointments that I would forget to make and remember[ing] to actually follow up with going to them. That’s huge. That sounds bad, but it’s huge. Little things down to the everyday activities … [such as] making sure that I took my vitamins. … So, that would be something I would forget. But with the journal, I’d open up [and realize], ‘Oh, yeah, I forgot. I need to do that.’ Even as little as brushing your teeth.”
PPJ helped break goals down into small steps	“Instead of saying, ‘I want to go to sober living,’ it’s, ‘I’m going to spend ten minutes today after dinner in the resource room researching sober living houses.’ I call it now small goals instead of big goals. I start with small goals, and I can keep building onto my small goals. If I say I want to move into sober living, it sounds like a way big[ger] order than if I start to break it down. [And] then I’m more likely to go and do it.”
PPJ created awareness of progress in recovery
Left side of PPJ promoted a feeling of improvement	“When you first get into a treatment center, you feel like … everything is off-balanced, and so when you see it balancing the other way, when the good things are outweighing the bad … it makes you feel like you can press forward more … [to] show people, show yourself and show whoever, that things are going better. It makes you feel much better about life.”
Right side of PPJ promoted a feeling of improvement	“I just noticed I would accomplish things and I would notice it when, if I wasn’t journaling or thinking about it, I wouldn’t have even thought about it. I just felt like I would evaluate my days and the things I was accomplishing and I would actually be able to sit in that happiness for a while, and be able to just see that I’m starting to get better. Sometimes I feel like I’m stuck and I’m not really moving forward; but in the journal, even just small things I was able to go back and see, ‘Yeah, look, I have accomplished that goal and I really didn’t even realize it. I’ve done that, I’ve improved that.’ And so it gave me a chance to actually see in writing how much I’ve gotten help here and how much my life has changed.”

Discussion

The 15 participants in this study registered 78% group attendance, completed 196 journal submissions, and achieved a grand mean of 39 bullet-point items per journal entry. In general, participants found PPJ to be easy, not difficult, pleasant, and satisfying. Over the course of the study, satisfaction with life, well-being, rewards from the environment, happiness with recovery, and confidence staying sober increased while depression decreased. Although there was no improvement in the other hypothesized mediators at the p < .05 threshold, effect sizes suggested improvement in optimism, craving, recovery capital, stress, and demoralization. This study did not have a comparison group, and changes occurred in the context of residential treatment for addiction; we cannot, therefore, conclude that PPJ caused the observed changes.

There was no change in negative affect, although negative affect was relatively low at baseline. Activated and unactivated positive affect were relatively high at baseline and small effect sizes suggested decreasing levels over the course of the study. It is noteworthy that we did not see improvement in these mood indicators, especially because so much previous work in positive psychology has shown this effect. A study of individuals in outpatient treatment for alcohol use disorders had shown decreases in negative affect and increases in unactivated positive affect owing to daily practice of the Three Good Things exercise (Krentzman et al., 2015). Because this exercise parallels PPJ’s Good things that happened prompt, it would be reasonable to expect improvement in mood associated with PPJ. It is possible that we did not register improvement in mood in the current study because of the lag sometimes experienced between participants’ PPJ writing and their completion of the Qualtrics questionnaires; outside of the three or four occasions when we handed out the iPads in group, participants visited the resident assistants’ office to borrow the iPads to complete the upload, which might have contributed to a lag between PPJ completion and assessment of mood. While both positive and negative affect can be fleeting (Fredrickson, 2013), even brief feelings of positive affect have been associated with lasting benefits, such as the expansion of psychosocial resources (Fredrickson, 2004a). However, affect might need to be captured immediately before it dissipates.

It seemed helpful that, in the current study, we applied Wong’s (2017) suggestions for learning gratitude in a group session. In The Three Good Things study with individuals with alcohol use disorders (Krentzman et al., 2015), many participants could not think of as many as three good things when listing them on their own as opposed to doing so in a group setting; in the current study, participants noted an average of six Good things that happened and 10 Things that I am grateful for.

Participants reported surprise that their good things “outweighed” their bad things, and this was the case despite our instruction to make Good things that happened and Things that I am grateful for lists as lengthy as possible but only to jot down what came to mind for the Bad things that happened list. Seeing the evidence of long lists of good things, written in their own hand, nonetheless seemed convincing to participants and provided them with a sense of optimism. The number of items listed under Good things that happened and Things that I am grateful for outnumbered the items listed under Bad things that happened by a ratio of more than 5:1.

Consistent with self-affirmation theory (Sherman & Hartson, 2011), Good things that happened and Things that I am grateful for seemed to strengthen participants and enervate their Bad things that happened list. While ours was a small developmental pilot study, our findings signal the potential for exercises that heighten awareness of what is going well during early recovery when negative thoughts and feelings are most prominent.

Our study has implications for the measurement of recovery. Researchers are increasingly less interested in substance use as a marker of recovery, focusing instead on factors related to well-being (Witkiewitz et al., 2020; Witkiewitz & Tucker, 2020)—even as the field has yet to settle on the specific aspects of well-being best-suited to measurement. Multiple hypothesized mediators we assessed had high ceiling effects at baseline, when our sample had, on average, 48 days of recovery. Trait gratitude was, on average, 37 at baseline (scale = 6–42), and commitment to sobriety was, on average, 29 at baseline (scale = 5–30), so we were not surprised that these factors did not increase over time. Commitment to sobriety decreased over the course of the study (with a small effect size), perhaps due to high scores for this factor at baseline or perhaps because participants felt more motivated earlier in their host-setting treatment episode, and slightly more discouraged, or even more realistic about their commitment, as time passed.

Other factors were comparatively lower at 48 days of recovery, such as satisfaction with life, with a mean of 15 (scale = 5–35), thus affording room to grow. It could be that some dimensions of well-being change rapidly with the onset of recovery, such as gratitude, while other characteristics are slower to change. Future research should continue to explore both which aspects of well-being change and the timing of change in early recovery.

Finally, this study makes a significant theoretical contribution by situating the relatively new body of literature on positive psychology and addiction recovery within the addiction field’s extensive tradition of reinforcement theory (Higgins et al., 2004). Positive psychological interventions can make life in recovery more positively reinforcing, and thus these literatures fit well in the larger context of reinforcement theories of addiction.

Limitations

Our results should be interpreted with some caution. It is important to emphasize that this study used a single-group research design; it did not employ a control group. This means that we do not know whether the improvement we observed was due to PPJ or due to the general improvement one might expect secondary to residential SUD treatment. The residential treatment host setting provided a range of evidence-based interventions, which also could have been responsible for positive change. Further, it is unclear how much benefit to attribute to PPJ and how much to attribute to common group therapy factors such as interpersonal learning and group cohesiveness (Yalom, 2005), although participants journaled on their own for the last 2 weeks of the study. Another limitation is this study’s reliance on a small sample. In addition, participants had to borrow iPads to complete their surveys (rather than using smartphones, which were prohibited at the site), creating occasional gaps of time between their completion of PPJ and questionnaire completion. On at least 14 occasions, PPJ entries were finished but technological challenges prohibited the completion of Qualtrics uploads. In residential treatment, there are significant restrictions to planning one’s upcoming day that might have limited the impact of PPJ. Since this was a developmental study, some procedures were modified cohort-to-cohort, such as the hour when the group would meet and the measures used. Also, the qualitative results are limited because we were unable to interview women who left the host site against staff advice or at staff request.

Future Directions

Future research should test PPJ in a well-powered randomized controlled trial. A larger study could determine for whom PPJ might work best. This was a small sample of women who were diverse demographically and clinically in terms of racial identity, number of children, substance most addicted to, length of sobriety, number of previous treatment episodes, and affiliation with mutual aid. It would be important in future work using larger samples to determine precisely for whom PPJ might be most helpful. For example, what is the experience of PPJ for people who identify as male or non-binary? Are all aspects of PPJ culturally valid across diverse racial and ethnic groups? Does PPJ work better for those with more severe addiction histories or for those with lower addiction severity? Is PPJ more or less effective for individuals with high rates of mutual-aid involvement, given that PPJ aligns with Step 10 and with the practice of gratitude common in such programs? At what stage in recovery is PPJ most effective? In addition to these questions, future research could also discern the effects of journaling in a group as opposed to independently. Finally, future research could ascertain how PPJ works as an adjunct to other recovery-supportive settings, such as outpatient treatment, sober living housing, or recovery community centers.

Conclusion

We found that PPJ was feasible and acceptable as an adjunctive intervention for 15 women in residential substance use disorder treatment. Aspects of positive psychology combined with aspects of behavioral activation reinforced what was going well in recovery, and PPJ activity co-occurred with improvement on several hypothesized mediators. PPJ fostered and amplified positive views of life in recovery that might counter the pervasive negativity in early recovery which has been both well documented in the literature and associated with substance use recurrence.

Availability of data and material:

We are unable to share a data file publicly; the small cohort size could make it possible to disclose the identity of research participants.